How to find out which indicators are the best to predict a trend

I have a table that the first column is a stock price sequence, and the second column is a tag that shows where the trend starts and ends, other columns are the value of many indicators (for example MACD, RSI, EMA, and so on), every indicator uses one column.

I want to use ML.NET to find out which indicators/indicators combination are most important to identify a trend (up/down trend).

I want to know if this can be resolved by ML.NET and which model I should be selected Binary classification, Multi-class classification, or others.